Hospice is considered a cost-effective and efficient way to deliver optimal end-of-life care. However, the use of hospice by terminally ill patients, especially among racial and ethnic minorities, is not proportional to the incidence of terminal illness in these populations. The proposed study will examine a cohort of patients who are referred to home hospice to identify patient, contextual and system factors that may affect hospice use. Of particular interest is the identification of the determinants of hospice use among referred racial and ethnic minorities. Three factors of interest are:1) patient factors, including age, sex, race/ethnicity, marital status, insurance, diagnosis, level of caregiver support, and need for services; 2) contextual factors, includingage, sex, race/ethnicity, marital status, education and employment levels of the population, per capita income, rate of health insurance coverage, ratios of physicians and hospital beds to the population and the morbidity and mortality rates for the top 10 hospice diagnoses in the patient's county of residence; and 3) system factors, including age, sex, race/ethnicity, education and training history, specialty, board certification status and hours spent in primary/clinical care for the patient's hospice referring/authorizing physician, and the effects of hospice's eligibility requirements on utilization. A major purpose is to develop and test a predictive model of hospice utilization derived from the Behavioral Model of Health Services Use and the Conceptual Model of Realized Access to End-of-Life Care to explain the variation in hospice use among referred patients as a function of the three types of factors. Specific aims are to: 1) identify patient factors, contextual factors of a patient's place of residence, and system factors associated with hospice use for all referred patients and for referred racial and ethnic minorities; 2) determine, for all referred patients and for referred racial and ethnic minorities, whether patient factors associated with hospice use are conditional on contextual or system factors; and 3) derive and test a predictive model for hospice use after referral.The data used to reach the goals will be extracted from the databases of two hospices, several publicy available sources, and the North Carolina Health Professions Data Book. SPSS 12.0 will be used to analyze the data using frequencies, correlations, probabilities and regression methods. The benefits of using hospice for terminally ill patients are well-supported in the literature from an economic, social and health care perspective. Understanding and improving disparities in hospice use will benefit public health by offering opportunities for all dying patients and their families to receive comprehensive and individualized care that promotes comfort, peace and dignity at the end-of-life.